QuantOM

About QuantOM
Members
Publications
Some projects
Seminars&conferences
What is OR ?
Podcasts
Links
Contact

 

Meetings 2015-2016

  • September 2016, Wednesday 14 (01:00 pm): by Johan Barthelemy (University of Wollongong)

    A stochastic activity-based model and a strategic agent framework for traffic simulation

    The work presented in this seminar focuses on the simulation of the mobility behaviour of (large) populations, i.e. the travel demand and its assignment on the road network.
    The travel demand is simulated using a stochastic activity-based approach in which the travel demand is derived from the activities performed by the individuals. The proposed model is distribution-based and requires only minimal information, but is designed to easily take advantage of any additional network-related data available. The proposed activity-based approach has been applied to the Belgian synthetic population. The quality of the agent behaviour is discussed using statistical criteria and results shows that the model produces satisfactory results.
    The demand can then be assigned on the network. For that purpose, a novel agent-based framework is being developed. The proposal relies on the assumption that travellers take routing policies rather than paths, leading us to introduce the possibility for each simulated agent to apply, in real time, a strategy allowing him to possibly re-route his path depending on the perceived local traffic conditions. The re-routing process allows the agents to directly react to any change in the road network. For the sake of simplicity, the strategy is modelled with a simple neural network whose parameters are determined during a preliminary training stage. The inputs of such neural network read the local information about the route network and the output gives the action to undertake: stay on the same path or modify it. As the agents use only local information, the overall network topology does not really matter, thus the strategy is able to cope with large networks. Numerical experiments are performed to test the robustness and adaptability to new environments. The methodology is also compared against MATSim and real world data. The outcome of the experiments suggest that this work-in-progress already produces encouraging results.

    Where: HEC-University of Liege - 14, rue Louvrex (N1)- 4000 Liège Etilux Room

  • May 2016, Wednesday 25 (10:30 am): Greet Vanden Berghe (KU Leuven)

    Theoretical and practical challenges in nurse rostering

    Health care is under high pressure to improve efficiency, given increasing requirements and decreasing resources. Among the activities to optimise, nurse rostering is one of the most relevant to address. The problem is computationally complex and has potential for improvement through automated decision support. Personnel rosters also have a considerable socio-economic impact.
    This optimisation problem has yielded ample practice-oriented operational research approaches. Despite the vast amount of academic research results, it remains hard for novice developers to profit from general insights or re-usable approaches. This `cold start’ issue can be partially explained by complicated regulations typical for personnel environments with 24/7 duties and different in almost every organisation. The very same issue also persists due to the lack of a theoretical framework for nurse rostering.
    From an academic point of view, interesting models have been developed for varying nurse rostering problems. The approaches range from self-rostering and manual problem decompositions to different levels of automated decision support.
    The seminar will focus on the challenging interplay between practical and theoretical nurse rostering contributions.

    Where: HEC-University of Liege - 14, rue Louvrex (N1)- 4000 Liège Room 1715

  • May 2016, Friday 20 (01:30 pm): Gilbert Laporte (HEC Montreal)

    The Fascinating History of the Vehicle Routing Problem

    The Vehicle Routing Problem (VRP), introduced in 1959 by Dantzig and Ramser, plays a central role in distribution management. It consists of designing a set least cost delivery or collection routes for a set of vehicles based at a depot and visiting a set of geographically scattered customers, subject to a variety of constraints. The most common constraints are capacity constraints, duration constraints and time windows. This talk will concentrate on the so-called classical VRP with capacity constraints only. The VRP is ubiquitous and highly important from an economic point of view. From a research perspective, it occupies a central role in operations research. Its study by the scientific community has fueled the development and growth of several families of exact and approximate algorithms. Exact algorithms such as branch-and-cut, column generation and branch-and-cut-and-price owe part of their evolution to the study of the VRP. Similarly, the most common classical heuristics and most of the more recent metaheuristics have been developed through the study of the VRP. In this talk I will highlight several of these developments. In spite of all the attention the VRP has received over the past 55 years, it can still only be solved exactly for relatively small instances (with slightly more than 100 customers) and the corresponding algorithms are rather intricate. Over the past 10 years or so, several powerful metaheuristics have been put forward for the approximate solutions of the VRP. The best ones combine concepts borrowed from local search and genetic search. Nowadays, the best metaheuristics can generate rather quickly solutions whose value lies within 1% of the best known solution values on a set of benchmark instances. This talk will also review these developments. It will close with some research outlooks.

    Where: HEC-University of Liege - 14, rue Louvrex (N1)- 4000 Liège Room 030

  • April 2016, Friday 29 (02:00 pm): Laurie Garrow (Georgia Institute of Technology)

    Estimation of discrete choice modeling parameters for revenue management models

    We develop a new parameter estimation routine for estimating multinomial logit models for which one alternative is completely censored. Our method is based on decomposing the log likelihood function into particular marginal and conditional components. Simulations based on industry hotel data clearly demonstrate the superior computational performance of our method over existing methods that are capable of estimating price effects. We consider extensions of our methodology to Generalized Extreme Value (GEV) discrete choice models that allow for flexible product substitution patterns. We show how the GEV choice-based sampling estimator can be applied to estimate parameters associated with censored alternatives, and derive identification rules that can be used to determine when these parameters are unique. Empirical examples based on simulated datasets demonstrate the large-sample consistency of estimators and provide insights into data requirements needed to estimate these models for finite samples.

    Where: HEC-University of Liege - 14, rue Louvrex (N1)- 4000 Liège Room 1715

  • April 2016, Wednesday 27 (10:30 am): Philippe Chevalier (UCL)

    Models for horizontal supply chain collaboration

    Horizontal collaboration can bring significant benefits to firms even in a competitive environment. In this talk we will present models for different types of collaboration ranging from Joint Ventures to joint planning of production equipment. Depending on the nature of the collaboration agreement, the incentive problems to be solved to reach an agreement that is both efficient and stable are different. In this talk we present different models to study collaboration agreements between firms. We analyse the investment decision, the capacity allocation decision and the short term planning decision.

    Where: HEC-University of Liege - 14, rue Louvrex (N1)- 4000 Liège Room Etilux

  • March 2016, Tuesday 8 (01:30 pm): Dr. Tom Van Lier (Vrije Universiteit Brussel)

    An external cost calculator framework for evaluating the sustainability of transport solutions

    Not taking external transport costs into account distorts market price signals, results in too high (road) traffic levels and avoids arriving at a societal optimum. The economic theory is sound, but practical applications are difficult due to the complexity of monetizing the different externalities in a sufficiently accurate way. The presentation will consist of three parts. A short introduction on the different transport externalities, with the current state-of-the-art methodologies for determining the external costs will be given. In a second part, some cases of external transport cost calculations for assessing the sustainability of transport options will be discussed, e.g in the field of horizontal logistics collaboration and teleworking. A third part will discuss the challenges for particular external cost categories in more detail, and will explore some potential ways forward to arrive at more accurate external cost assessments.

    Where: HEC-University of Liege - 14, rue Louvrex (N1)- 4000 Liège Room 224
    Presentation

  • February 2016, Wednesday 17 (10:30 am): Thierry Vanelslander (University of Antwerp)

    PORT-RELATED INNOVATION: IS THERE A RECIPE FOR SUCCESS?

    Innovation in general seems to happen very rapidly these days. However, the poor innovative strength displayed by the transport sector in the broad sense often contrasts strongly with that evidenced elsewhere. At the same time, it can be concluded from existing literature and studies that quite a lot of innovative concepts in transportation have been studied in detail. The main focus hitherto however has always been on inventing or introducing new concepts and procedures.
    First, a typology of different types of innovation is being developed, illustrated with the actual occurrence in practice of each innovation type.
    This paper assesses which innovations will generate which chain impacts, what conditions will conduce actors to innovate, or prevent them from doing so, and finally also what governments can do to stimulate innovation.
    The present paper makes an application to innovation by seaport-related operators through a number of cases. On these cases, a standard set of information is collected. For collecting the information, use is made of existing documents on the cases at hand, on which a scientific review is performed. Equally, sector and project contacts are used for verifying and completing the information. In order to be able to derive meaningful conclusions from the set of cases, the fuzzy-set qualitative comparative analysis (fsQCA) technique is used.
    The analysis shows that no unique recipe for innovation success exists, but that sub-groups of cases can be discerned which have common characteristics that consistently lead to innovation success, dealing with the terminal operators, the shipping lines, infrastructure and the innovation champion.
    The findings are particularly relevant for those in business and in policy-making who are involved in implementing or promoting innovation.

    Where: HEC-University of Liege - 14, rue Louvrex (N1)- 4000 Liège Room Etilux

  • February 2016, Tuesday 1 (01:30 pm): by Diego Cattaruzza (Ecole Centrale de Lille)

    Vehicle Routing Problems for City Logistics

    City Logistics has been introduced to study and understand the transport activities in urban areas. Optimisation of urban delivery planning needs to take into account the dynamics of the city in order to provide efficient services. In nowadays delivery systems it is common to forbid heavy trucks to enter city centers. Trucks are forced to unload at logistic platforms located in the outskirts of the city. Merchandise is then loaded into smaller (and possibly eco-friendly) vehicles in charge of final delivery to customers.
    First we present a survey of the movements of goods that occur in cities. This motivates the study we conducted and the development of optimisation tools for delivery activities in urban areas.
    Second, we introduce the Multi-Trip Vehicle Routing Problem with Time Windows and Release Dates.
    Routing problems that take into account time windows and multiple delivery trips are commonly studied by the scientific community, especially when applied to the urban context. On the other side, the study of routing problems that consider release dates on the products is not. Here the release date models the instant the merchandise is available at the logistic platform and ready for final dispatching. A genetic procedure is presented to tackle the problem.
    This work was conducted in the context of the MODUM project (http://www-lipn.univ-paris13.fr/modum) that was founded by the French National Research Agency - ANR.

    Where: HEC-University of Liege - 14, rue Louvrex (N1)- 4000 Liège Room Etilux

  • December 2015, Thursday 3 (11:00 am): by Pieter Vansteenwegen (KU Leuven)

    Advanced methods for robust timetabling for the Belgian railway network

    In nearly saturated station areas the limited capacity is one of the main reasons of delay propagation. Spreading the trains well in these areas has a big impact on the total travel time in practice of all passengers in the railway network in case of frequently occurring small delays. We focus on improving the performance in the bottleneck of the network in order to improve the performance of the whole railway network. We propose a heuristic method to improve an existing timetable and also a method that builds from scratch a routing plan and a cyclic timetable that optimally spread the trains in space and time. These methods are applied to a case study of the Brussels' railway station area, the bottleneck in the Belgian railway system. Both methods are able to improve the timetabling significantly.

    Where: HEC-University of Liege - 14, rue Louvrex (N1)- 4000 Liège Room 138

  • November 2015, Friday 6 (01:30 pm): by Roel Leus (KU Leuven)

    An exact algorithm for parallel machine scheduling with conflicts

    We consider an extension of classic parallel machine scheduling, where an undirected conflict graph is part of the input. Each node in the graph represents a job and an edge implies that its two jobs cannot be scheduled on the same machine. The goal is to find an assignment of the jobs to the machines such that the maximum completion time is minimized. We present an exact algorithm based on branch and price.

    Where: HEC-University of Liege - 14, rue Louvrex (N1)- 4000 Liège Room 034

  • October 2015, Friday 9 (01:30 pm): by Professor Christoph Buchheim (TU Dortmund)

    Robust Preprocessing for Real-time Combinatorial Optimization

    We investigate a new concept in robust combinatorial optimization that aims at alleviating the so-called price of robustness: instead of a single solution as in the strictly robust setting, we allow to compute k solutions for a given number k, assuming that we may choose the best one of these solutions once the actual scenario is revealed. The aim is thus to solve a min-max-min problem: find k feasible solutions such that the respective best of them is optimal in the worst case. In a typical application of such an approach, an uncertain combinatorial optimization problem needs to be solved frequently and quickly, e.g., a parcel service needs to solve the same vehicle routing problem every day under changing traffic situations. After a potentially expensive preprocessing phase for computing the k solutions, the best of them can then be chosen very quickly every day once the actual traffic situation is known. We discuss the complexity of the resulting min-max-min problem for convex and (time permitting) discrete uncertainty. In the case of a convex uncertainty set, we show that the latter problem is as easy as the underlying certain problem if k is greater than the number of variables and if we can optimize a linear function over the uncertainty set in polynomial time. In the case of discrete uncertainty with a fixed number of scenarios, we show, among other results, that the problem is weakly NP-hard and admits an FPTAS for the shortest path problem, the minimum spanning tree problem, and the knapsack problem, while it is tractable for the min-cut problem.
    (joint work with Jannis Kurtz)

    Where: HEC-University of Liege - 14, rue Louvrex (N1)- 4000 Liège Room Etilux


 
 
  HEC-Management School - University of Liège